Chromatography columns have been extensively developed and are used routinely in both analytical and preparative chromatography. As is well known, the separation in a chromatography column of a sample (also termed an analyte or solute) comprising a mixture of components is achieved by dissolving the sample in an eluant to form a fluid mobile phase and passing the mobile phase through a stationary phase typically packed within a tubular column, thereby causing the sample to separate into its components due to the differences in the partitioning between the mobile and stationary phases of the different components (i.e. the components have different partition coefficients). The eluant fluid is most commonly a liquid but may be another fluid such as a supercritical fluid (SCF) and the invention relates to columns used with liquid or SCF mobile phases. The invention, however, does not relate to gas chromatography (i.e. the fluid is not a gas). In column chromatography the stationary phase is typically in the form of a bed of packed particles or a porous monolithic block within a column. This invention relates especially to packed columns but it is not limited to only packed columns. Often the columns comprise reusable columns with disposable cartridges, both of which are usually cylindrical. This invention may be used with cylindrical columns (most preferably circular cylindrical columns) or columns having other cross-sectional profiles. That is, the wall of the column may have numerous cross-sectional profiles but most preferably has a circular profile in its transverse cross section.
As described, the mobile phase is passed through the column and the eluate leaving the column is detected as a function of time. The detected signal variation with time, the chromatogram, indicates the presence of different components within the mixture. The degree of separation of the different components depends upon the separation efficiency or resolution of the column. The resolution of the column depends upon many factors. Such factors include the nature of the mobile and stationary phases, which have been extensively studied and developed.
It is also known that the shape of the injected sample is a factor affecting the resolution of the column. The sample is usually introduced into the column substantially across the diameter of the column. Head fittings such as frits are usually used at the inlet of the column. Distributors may be used to aid in distributing the sample to try and obtain a uniform layer of sample across the whole diameter of the column. U.S. Pat. No. 4,999,102 describes a manifold system for evenly distributing liquid to, and/or evenly collecting a liquid from, a cell of a large scale separator system. Where used as an inlet, the manifold distributes a single stream into a plurality of streams through a tier of branches and the final multiplicity of fluid passage devices is arranged in a pattern which assures approximately even distribution of liquid across the entire cell at the interface between the manifold and the cell. Where used as an outlet the manifold is arranged in reverse to collect from an even distribution across the cell and, via a tier of branches, join the streams into a single conduit. In a similar manner, U.S. Pat. No. 5,124,133 describes a system for a uniform flow profile of a liquid through a packed bed that involves distributing the liquid evenly across the top surface of the packed bed.
Other types of inlet system are known, often for different purposes. For example, EP 371 648 A discloses an apparatus designed for displacement chromatography in which an inlet distribution manifold is present. Each liquid in the displacement chromatography sequence flows through the same radial regions of the column. On the other hand, JP 62-063857 A describes a plate-shaped column having a plurality of inlets for possible multiple samples in which an electric field may be applied in an orthogonal direction to the inlet flow to provide a two dimensional separation across the column plate, which also has a plurality of outlets.
In JP 62-240857 A is described a chromatography column for preparative chromatography in which there is provided annular separation of flow exiting the column such that central flow streams can be isolated from peripheral flow streams. The peak shape of the eluting bands from the central rings were better compared to those of the eluting bands from the outer rings, which showed significant band distortion.
EP 257 582 describes a preparative scale chromatography column in which sample is loaded into the column through a tube inserted into the bed from the column outlet. This tube is centered in the bed approximately 20% below the column inlet. That is, there is packing material above the tube. Once the sample is loaded onto the column a mobile phase elutes the sample through the column (outside of the sample introduction tube). Sample then exits the outlet of the column through a series of holes or slots at the outlet. The outlet has slots at concentric locations, but is without a central exit slot since the introduction tube is in the central section of the column.
More recently, in relation to high performance liquid chromatography (HPLC), Shalliker et al. in the late 1990s enabled fluid flow in a column to be visualised. In glass columns, using a stationary phase and a mobile phase with the same refractive index, they were able to visualise fluid flow with the aid of dye markers. Their results showed that the manner in which sample is introduced to the bed of a chromatography column is important. Ideally the injection plug should be a cylindrical-square plug, but various factors ultimately lead to a parabolic sample band, for example, as the frit porosity decreased plug flow becomes more parabolic. Broyles, Shalliker and Guiochon, (‘Visualization of Solute Migration in Chromatographic Columns. Influence of the Frit Porosity’. J. Chromatography A., 917 (2001) 1-22) illustrated that as the inlet frit porosity decreased plug flow became more parabolic. Use of a distributor improved the uniformity of the flow velocity across the radial cross section of the bed, however, at the cost increasing the axial dispersion, which ultimately led to a substantial decrease in separation performance. Irrespective of the frit porosity or whether or not a distributor was employed, sample distribution was not uniform across the column radial cross section. Shalliker et. al. (J. Chromatography A, 865 (1999) 83-95) showed that neither frits nor distributors serve to distribute sample uniformly across the column, and there is a higher tendency for the sample to be more concentrated in the central region of the bed (conversely, more dilute in the perimeter region of the bed). Furthermore, the frit diameter should match that of the column internal diameter (Broyles, Shalliker and Guiochon, J. Chromatography A, 855 (1999), 367-382). If not, with a frit having a narrower diameter than that of the column, very serious parabolic flow results. It may be intuitive to the chromatographer that the frit diameter must equal that of the column internal diameter, but it is not always possible to achieve, particularly in the case of columns prepared by axial compression where the head fitting is inside the column itself. For these types of columns the diameter of frit must be less than the column as the frit must be contained within a housing unit in order to prevent leakage and damage to the column wall.
In another study, Shalliker et al (‘Physical Evidence of two Wall Effects in Liquid Chromatography’. J. Chromatography A. 888 (2000) 1-12 and ‘Visualization of Solute Migration in Liquid Chromatography Columns’. J. Chromatography A., 826 (1998) 1-13) demonstrated that when sample was introduced into the bed at a depth approximately 1 cm below the frit, via a central point injection (CPI), migration of the sample was coincident within the central radial section of the bed. While excellent separation efficiency was observed by Shalliker et. al. when utilising the CPI technique, it should be pointed out that such a sample introduction method is tedious, and prone to be destructive to the column bed. Hence the technique is not suited to routine applications. Furthermore, whilst central point injections allow for the most efficient means of solute transport along the column since such injections allow the sample to be concentrated in a local, central zone within the most efficiently packed region of the column, modern columns have effectively prevented the use of such injection techniques, since valve injection disperses sample across the column as it enters by a frit and distributor at the top of the column. As such, central point injection processes have been largely abandoned.
The effect of the presence of the column wall and the packing of the column near to the walls on the flow of sample has been discussed. Knox, Laird and Raven, (J. Chromatography, 122 (1976), 129-145), showed that flow velocity very close to the wall was somewhat higher than in the centre of the column. They also showed that a disturbed region of column packing extends into the column causing serious band broadening and peak distortion. The wall region in an otherwise well packed LC column may extend about 30 particle diameters into the column. Shalliker, Broyles and Guiochon, (J. Chromatography A, 888 (2000) 1-12) described the presence of two wall effects. Both wall effects are caused by heterogeneity in the column packing. A particle packed chromatography column has a lower bed density in the central (radial) section of the column, the central section being relatively uniform, but beyond this zone packing density gradually increases towards the wall. Hence flow velocity is highest in the central section of the column and lowest near the wall, and this contributes to a parabolic plug flow. However, in the immediate vicinity of the wall the packing density rapidly decreases. This factor is a result of the geometrical nature of the column and the particles. Both the particles and the column are rigid; neither can distort to accommodate the other. Hence the void space increases at the wall, and it is here that the column permeability is at its highest. Hence the flow velocity is greatest in the region immediately adjacent to the column wall. Both these wall effects contribute greatly to decreasing the efficiency of migration.
Broyles, Shalliker and Guiochon, (J. Chromatography A, 867 (2000), 71-92), demonstrated that, amongst other things, the migration distance after a given time varies markedly with radial position and that the exit fitting affects the peak shape detected post-column.
It is therefore known that sample migration through a particle packed chromatography column is not described by a cylindrical plug flow model. Numerous factors, including the nature of the frit, the presence of a distributor, the injection process itself, the wall effect and the heterogeneity of the packing density, lead to a sample plug or band that is parabolic or bowl-like in shape, generally having a higher concentration central region that is moving at a higher velocity than a more dilute region nearer the wall. There is also a region of low concentration fast flow very close to the wall. The generally parabolic shape of the band places greater demand on the efficiency of the column. More plates are needed to separate parabolically broadened zones than cylindrically broadened zones. Hence, longer columns are necessary to effect separation, which results in an increase in time, possibly resulting also in a decrease in the potential flow velocity in order to accommodate a longer column. Monolithic stationary phases in columns may also suffer from wall effects and plug flow through such columns is again not cylindrical.
Against this background the present invention has been made.